Papua New Guinean Genomes Reveal the Complex Settlement of North Sahul

Abstract

The settlement of Sahul, the lost continent of Oceania, remains one of the most ancient and debated human migrations. Modern New Guineans inherited a unique genetic diversity tracing back 50,000 years, and yet there is currently no model reconstructing their past population dynamics. We generated 58 new whole-genome sequences from Papua New Guinea, filling geographical gaps in previous sampling, specifically to address alternative scenarios of the initial migration to Sahul and the settlement of New Guinea. Here, we present the first genomic models for the settlement of northeast Sahul considering one or two migrations from Wallacea. Both models fit our data set, reinforcing the idea that ancestral groups to New Guinean and Indigenous Australians split early, potentially during their migration in Wallacea where the northern route could have been favored. The earliest period of human presence in Sahul was an era of interactions and gene flow between related but already differentiated groups, from whom all modern New Guineans, Bismarck islanders, and Indigenous Australians descend. The settlement of New Guinea was probably initiated from its southeast region, where the oldest archaeological sites have been found. This was followed by two migrations into the south and north lowlands that ultimately reached the west and east highlands. We also identify ancient gene flows between populations in New Guinea, Australia, East Indonesia, and the Bismarck Archipelago, emphasizing the fact that the anthropological landscape during the early period of Sahul settlement was highly dynamic rather than the traditional view of extensive isolation.

Keywords: Oceania; Papuan; Sahul; demographic history; human genome.

Fig. 1.

Genetic diversity of individuals from New Guinea, Wallacea, the Bismarck Archipelago, and northeast Australia. (A) Geographic locations of all 58 genetic subgroups defined by the fineSTRUCTURE analysis. (B) Principal component analysis representing 26.7% of the total variability of all individual genomes from the 58 subgroups, masked for Asian genetic ancestry. The two first components are represented (PC1 and PC2). (C) ADMIXTURE plot for eight genetic components summarized by genetic subgroups (based on K = 8 from supplementary fig. 7, Supplementary Material online). The gray component combines four components mainly present in African and Eurasian groups. Numbers above bar plots identify genetic subgroups referred in supplementary table 2, Supplementary Material online. (D) TREEMIX plot for all 58 subgroups (based on supplementary fig. 7, Supplementary Material online). All four panels used the same color scheme to identify the 58 genetic subgroups.

Fig. 2.

Scenario of the settlement of Sahul. (A) f3-outgroup analysis for all Sahul groups compared with populations from the North Maluku Islands (y axis) or Lesser Sunda Islands (x axis). Populations are identified by the same color scheme as in figure 1. (B) Best fitting qpGraph models considering one migration to Sahul (model A; Z score=−2.570) or two migrations to Sahul (model B; Z score = 2.301). One subgroup was used as representant of a region: populations from North Maluku (Wallacea subgroup 23, following fig. 1C and defined in supplementary table 2, Supplementary Material online), Lesser Sunda Islands (Wallacea subgroup 13), New Guinea (New Guinea subgroup 16), the Bismarck Archipelago (Bismarck subgroup 5), northeast Australia (Australia subgroup 1), and Africa (Africa_5). Solid arrows are labeled with branch lengths in 1,000 times drift units. Dotted arrows indicate gene flows whose proportions are indicated by percentages. (C) Relative cross-coalescence rate curves estimated by MSMC2 between populations from Wallacea (Lesser Sunda Islands, Wallacea subgroup 13), New Guinea (New Guinea subgroup 16), the Bismarck Archipelago (Bismarck subgroup 5), and northeast Australia (Australia subgroup 1).

Fig. 3.

Scenario of the settlement of New Guinea. (A) Best fitting qpGraph model for populations from New Guinean southeast lowlands (New Guinea subgroup 16, following fig. 1C and defined in supplementary table 2, Supplementary Material online), north lowlands (New Guinea subgroup 28), south lowlands (New Guinea subgroup 39), east highlands (New Guinea subgroup 45), west highlands (New Guinea subgroup 55), and Africa (Africa_5) (Z score = 2.759). Solid arrows are labeled with branch lengths in 1,000 times drift units. Dotted arrows indicate gene flows whose proportions are indicated by percentages. (B) Relative cross-coalescence rate curves estimated by MSMC2 between populations from New Guinean lowlands and highlands.

Fig. 4.

Genomic scenario of the human dynamics in north Sahul during the Upper Pleistocene. Light blue arrows represent the settlement of north Sahul following a scenario with one migration to Sahul (model A). The dark blue arrows represent the settlement of north Sahul following a scenario with two migrations to Sahul (model B). The dotted dark blue arrow represents a gene flow following a scenario with two migrations to Sahul (model B). Yellow arrows represent the settlement of New Guinea. The dotted orange arrows represent gene flows after the initial settlement of Sahul. The black dots represent approximate locations of archaeological sites older than 40 ka (O’Connor 2007; Summerhayes et al. 2017). The dashed black lines represent ecological lines in Wallacea. The dark gray areas represent current land area, with light gray areas representing estimated land area around 50,000 years ago based on paleogeological reconstructions (Coller 2009).